Remove extraneous data

README.md

@@ -5,4 +5,6 @@ language:
 
 # JBLiMP
 
-This is the data from "JBLiMP: Japanese Benchmark of Linguistic Minimal Pairs" (Someya and Oseki, 2023). For details see [the original git repo](https://github.com/osekilab/JBLiMP) or [the paper](https://aclanthology.org/2023.findings-eacl.117/). 
+This is the data from "JBLiMP: Japanese Benchmark of Linguistic Minimal Pairs" (Someya and Oseki, 2023). Only the validated pairs used for benchmarks are included, and only in JSONL format, since it's redundant with the TSV. 
+
+For details see [the original git repo](https://github.com/osekilab/JBLiMP) or [the paper](https://aclanthology.org/2023.findings-eacl.117/).